An Interview with
NFL, NCAA Football
YL: When did you develop an interest in rating team
strength and what caused this interest?
Herman: I've always enjoyed the game of football and was
a fan of Tennessee back into the mid-Thirties. My interest in ratings came
during the 1944 and 1945 seasons primarily as I enjoyed the Litkenhous ratings
in the Knoxville News-Sentinel. Later, I learned of the Dunkel Index and the
Williamson System from other papers. I was so fascinated by these various
ratings that I wanted to try my own hand at it.
My first simple system was developed in 1946 and was improved
as I finished high school in 1948. I toyed with this matter all the way through
a period of college and while in the service during the Korean Conflict.
Upon my return home in 1953 I began publishing that system in
the local Middlesboro (KY) Daily News. I met Dr. Litkenhous while in Graduate
School at Vanderbilt University. We corresponded from that time until his death.
In 1965 I got interested in linear programming (but I am far
from an expert in that field) and saw how I could use some ideas there to
develop a better system. I first published this system in 1966. I call it the
Matthews Grid Ratings. NCAA and NFL football are both rated at present and have
been for almost every year since 1966.
I do not attempt basketball. There are too many teams and
games, and I do not have the drive (old age creeping on, I suppose) to attempt
such an added workload, particularly since my duties at Lincoln Memorial University (Harrogate, TN)
must come first.
YL: That's very interesting. You've seen quite a change
in the calculation and presentation of football ratings. The first ratings I
ever saw that impressed me were Harmon's and Jeff Sagarin's whose widely
publicized ratings escalated my interest in the numbers. What do you think are
the most important factors in calculating the team strengths?
Herman: Most certainly some near-correct balance between
the consistency of rankings (winners placed over losers to a high degree) and
some kind of average of differences in scores.
Scores certainly cannot be ignored. If A and B play the same
teams and A wins on average by 30 points more than B, then A most certainly is
the stronger team. If, for example, as has been suggested, the score differences
are truncated at 21 points and B wins every game by 21 points and A wins by 51
points, then the truncation will show A = B in rankings or ratings--simply
My system does a good job of finding this balance. The
predictions are usually within a couple of percents of the best predictions in a
given season, while the consistency is also very good. Have you seen--a dumb
question, I guess--the site maintained by David Wilson? At the bottom of the
rankings is a line called the Ranking Violation %. Only two systems in the list
(of 55 systems) rank higher in consistency than mine.
Yes, and I have exchanged ratings with several of the old
timers, some for many years: Litkenhous, Dunkel (Dick Jr.), Williamson (and son
Mitch after his death), DeVold (of the old Football News--he was down to see me
last summer), and Dick Poling (who used to visit our campus at Lincoln Memorial
University and said it was the most beautiful in the nation, adding "and
I've seen them all"). In addition, two of my former mathematics majors rate
high school teams for their states in an official capacity and another
well-known statistician used to do a college scouting report similar to those so
popular in yesterday's and today's paper.
YL: Having the history in rating teams, do you see the
Internet as good or as bad regarding computer ratings? Do you think the
publication of so many ratings could hurt the 'business'?
Herman: The presence of so many is neither good nor bad.
It certainly gives one more information and reveals how diverse the thinking is
regarding this matter. I don't think it hurts "business."
Perhaps someone(s) will come up with a better system than
many presently being published or offered there. Everyone has this right and, if
so inclined, should participate.
YL: In your opinion, where do we go from here? What role
will computer ratings have in determining the bowls or playoffs? Will they ever
replace human polls?
Herman: There apparently will be little change until the
present BCS "contract" is completed. I hope computer ratings continue
to play an important part in determining the bowls or playoffs if playoffs are
begun in the future. Talk of 32 or even 64 teams in a playoff is neither
responsible nor practical. I can't imagine any system or poll as having much
validity unless the top team is identifiable among the top eight teams.
Personally, if a playoff is ever begun, I would think only
four teams would be required. Neither polls nor computer ratings will replace
the other because of their independence from each other. Neither needs the other
to exist. Besides, think of the sheer pleasure derived from the varieties of
opinions expressed in polls or computer ratings.
YL: What is your profession?
Herman: I am an Associate Professor of Mathematics and
presently the Chairman of the Department of Mathematics and Computer Science at
Lincoln Memorial University, Harrogate, Tennessee.
YL: Is Harrogate the place you call home?
Herman: No. I live in Middlesboro, Kentucky, just across
the state line from Harrogate. We are in the KY-TN-VA corner at Cumberland Gap
where Daniel Boone and his party entered Kentucky the first time.
Middlesboro has its share of great football players, I should
add. One was "Ig" Fuson of the Tennessee team in 1940 that played
Boston College. His brother "Ug" played on the great Army team with
Davis and Blanchard. Ig was killed in Germany in WWII and Ug was wounded in
Korea and died not too many months after returning home. Another player was
Leonard Coffman who played on the great 1938 Tennessee team. Still another is a
fellow known to TV fans as Lee Majors (Harvey Lee Yeary). Finally, one of our
deceased superintendents of the school system played on the Centre team that
upset Harvard. Middlesboro is quite a football town.
Sorry for the long answer. I just couldn't pass up this
YL: That's fine. After all, it's your interview and you
can say whatever you wish. I'm sure you'll bring some good memories to longtime
fans. What other, if any, sports and teams do you particularly like?
Herman: I enjoy college basketball, but not with the
intensity of football. I'm somewhat passive about nearly all other sports. The
child in me still loves marbles!
YL: What else do we need to know about Herman Matthews?
Herman: I have had a very good life, a good family, lots
of friends, and an immense joy in teaching so many years. There are so many
things for which to be thankful!
We would like to thank Mr. Matthews for
this interview. His web site can be found at http://www.expertpicks.com/